STEM newsletter

Modelling enterprise IT infrastructure

30 April 2004

As the current economic climate brings more focus on operating costs, businesses in all sectors must continuously tune their IT infrastructure to help drive cost savings, while at the same time ensuring operational excellence. Dasein Ltd in Ireland has been using STEM to develop business-planning models for enterprise clients in EMEA and North America, enabling them to assess the wide range of options now available for their IT infrastructure.

How can businesses manage their way forward?

The shape of enterprise IT infrastructure is changing rapidly as storage requirements and staff costs continue to increase, while the costs of communications and hardware decrease (as illustrated below). Its future shape is likely to be determined by developments in ‘utility computing’, greater use of the Internet for all teleservices, off-premises storage, teleworking, mobile services and outsourcing. Also, increased regulatory requirements resulting from Basel II and Sarbanes Oxley will drive the need for best practice in terms of security and availability.

Evolution of costs over time

Managers trying to plan their company’s IT investment are faced with a number of pressing issues, such as:

  • understanding the costs and benefits of business continuity and security
  • deciding on different roll-out scenarios with respect to technology, timeline, vendor, and which sourcing strategy should be adopted
  • tackling the misunderstandings that can occur between IT, Finance and business-unit heads when deciding on a solution
  • creating a clearer picture of what may be an increasingly complex infrastructure despite limited staff and tighter budgets.

These issues are non-trivial, and a high level of planning and communication is required to ensure success. This is where a planning tool such as STEM can help. STEM’s data structure is able to capture the architecture of any IT infrastructure and model its financial and resource dynamics over time. It addition, it can act as a communications bridgehead between the technical specialists proposing a solution and those who have to finance it.

STEM for the enterprise

If you can define an IT infrastructure in terms of its resources, the demands made upon it by a user population, and its financial parameters in terms of capital outlay, operational and maintenance costs, etc., you can model it in STEM. Typically the enterprise infrastructure will comprise:

  • desktop or laptop PCs, phone (fixed and mobile), LAN connection and all client software
  • LAN or WAN for all required telecoms services (voice, data, Web connectivity, etc.)
  • servers and other concentrated areas of computing resource and associated software
  • storage, backup and archival
  • provision of environment, power, etc., to adequate service levels
  • staff.

The following is a trivial client-server model.

We have a user demand for storage (USERS) that will grow over time. This demand has a number of services (illustrated here by Email, File and Print Services) which can be modelled to service an increasing proportion of the storage demand and which in turn can be used to drive a demand for disk resources (Disk). We can set the initial disk capacity to be proportional to the number of CPUs already deployed (CPU count) to provide a starting point for storage provision. As disk resources grow to meet demand over time, backup requirements can be calculated using the Archiving Req. transformation to calculate backup requirements. Capital, operational and other costs can be inputted for each resource, along with such features as specifying the financial versus physical life of the resource, setting year 1 maintenance to be zero for warranties, etc.

When the model is run, we can see how resources increase with time and how costs vary both in terms of operational expenditure and capital expenditure. We can factor in time-varying parameters for increasing costs for operations (staff costs may be a consideration here) and decreasing cost of hardware and examine how these can affect overall expenditure, to produce a net present value analysis. We can also see how costs are allocated to the services provided, to identify where most of the investment is being dissipated.

Resource capital expenditure

 

Resource operating charges

 

Resource installed and incremental units

Now that we have created the model, we can start to examine other possibilities. Supposing we want to analyse the possible outsourcing of the storage requirement, an option that would eliminate the capital component and increase the associated operational cost. We can model an elimination of the backup requirements, as we assume our outsource arrangement will handle this (on a per-Megabyte basis), and set the model so we are adding extra disk resource on a smaller, incremental basis. This provides a mechanism for comparing the real cost of in-house versus off-premises storage. We do not change the overall shape of the model, we just change the financial parameters and the parameters that define when extra resources are needed.

With a flexible model of the enterprise’s IT infrastructure in place, planning and decision-making cycles can become more frequent because they are less costly and time-consuming. Once the overall architecture is agreed, ongoing reviews of technology acquisition, cost allocation and vendor services can be made more easily and made visible to all.

Further applications

Other examples of where modelling can be brought to bear include:

  • Migration to a thin-client architecture that has a centralised or utility-based server environment. We could take the situation where the initial environment comprises classic desktop PC clients and branch or local-level servers. Typically such a migration has to be done in situ while the thin client is gradually rolled out. The company would also need to dynamically upgrade its WAN because bandwidth requirements start to increase as more thin clients are deployed. One could model different scenarios where one uses varying speeds of roll-out, or decide to outsource (or not) the centralised server environment and WAN.
  • Desktop upgrades. Similar to the thin client in concept except we are looking at a full change-out of the desktop, possibly with new storage, LAN and WAN requirements.
  • Migration of legacy WAN to VPN and migration to VoIP for a branch network. e.g. retail outlets, government departments, financial institutions.
  • Deployment of a business-to-business network which forms a supply chain for a business which relies on heavily outsourced manufacturing, logistics and fulfilment activities. Here we can look at options where the company buys its own VPN and associated security, versus having a third party provide the full mediation service.
  • Migration of IT services provided in-house to a fully outsourced model, or to somewhere in between. How much does the business need to own, as against buying on an as-needed basis?
  • Modelling a CRM in terms of the mix of fulfilment channels, be they call centre, Web, mail-based, etc. Again we are aiming to dimension the service resources required to meet demand and determine whether it is be better to outsource or own the associated infrastructure and its operation.
  • Vendor option analysis – comparison of different vendor offerings to deliver a given requirement
  • Business continuity and disaster recovery. Should the company rent this as an extra standby facility, or structure its IT infrastructure to provide similar resilience?

Conclusion

Irrespective of what is modelled, STEM’s main aim is to provide speed of analysis, clear understanding of what is happening over time, good communication between all affected parties and a good basis for decision-making. Businesses face a bewildering set of options for their IT infrastructure, and competitiveness requires good planning and decision-making in this area. Mistakes are very easy to make, given that there is less and less time available and fewer slack IT resources. Such mistakes can result in downtime, lost customers or worse.


Dasein Limited is a technology and business management consultancy dedicated to maximising the opportunities available in the ICT space from both an enterprise and service provider perspective.

© Implied Logic Limited